AI platform Midjourney bans journalist who created Trump's fake images of getting arrested

Yesterday, about 50 AI-generated pictures envisioning former President Donald Trump’s arrest circulated across Twitter.  The images caused confusion for some users because the Midjourney v5 engine used to make them usually produces very realistic-looking content. The Midjourney user behind the pictures, Bellingcat creator Eliot Higgins, claimed to Buzzfeed News that he was barred from Midjourney because of the images. Buzzfeed also noted that Midjourney has prohibited the use of the term “arrest” in text prompts to create images. User suspended over Trump’s pictures The suspension has not yet been verified by Midjourney. According to an Ars investigation, prompts containing Donald Trump’s name appear to be banned on the engine, resulting in an error notice stating that the prompt was stopped because “it may conflict with our content policy."   Other users reportedly received the same error notice when they used prompts that included other Trump family members as well as all of the leaders’ names.   **Also read: AI-generated pictures of former President Trump getting arrested send Twitter into a frenzy** Higgins told Buzzfeed that he knows why Midjourney may have banned him, claiming that making the popular Twitter post was “pushing my luck.” Higgins told Ars that while the pictures have received mostly favourable reactions, he is also receiving “quite a few angry” Trump supporters. “I had no intention of doing any clever criticism or anything like that,” Higgins told Buzzfeed, adding that he was simply having fun envisioning Trump’s arrest with the pictures. “But then it kind of took on its own life." “What really stood out,” Higgins continued, was “the number of people who don’t seem to be aware of Midjourney’s existence or what it’s capable of.” He believes that creating the images “had a positive impact in raising awareness of what types of images can now be faked,” and that AI platforms such as Midjourney “probably shouldn’t train their datasets on people they don’t want re-created." Twitter’s regulation for AI-generated images Higgins told Wired yesterday how Twitter users could escape being duped by convincing AI-generated pictures such as the phoney Trump arrest photos. Celebrities, according to Higgins, are easy targets for deepfakes because “it’s clear that the more famous a person is, the more images the AI has had to learn from." “So very famous people are rendered extremely well, whereas less famous people are typically a little wonky,” Higgins explained to Wired. **Also read: Twitter to label deepfakes, deceptive media leading up to 2020 US Presidential elections** Higgins’ pictures on Twitter are now labelled with a community comment indicating that they are fake. According to Twitter’s policy, users “may not deceptively share synthetic or manipulated media that is likely to cause harm,” and these marks “help people understand their authenticity” and “provide additional context." Higgins’ full thread can still be found on Twitter, where some pictures have been watched 4.8 million times as of this writing. The entire Trump case While Trump earlier stated that his indictment would take place yesterday, this did not materialise. MSNBC and The New York Times are now saying that he could be arrested in the coming days for secret money payments to porn actress Stormy Daniels. **Also read: Stormy Daniels: The porn star who has become a thorn in Donald Trump's side** While Ars has found no Trump remarks on Higgins’ AI-generated pictures, Forbes reported that Trump has told advisers that if he is arrested, he wants it to be a “spectacle” that involves a “perp walk” in shackles to avoid appearing feeble to his followers. Read all the  Latest News ,  Trending News ,  Cricket News ,  Bollywood News , India News  and  Entertainment News  here. Follow us on  Facebook,  Twitter and  Instagram.